ABSTRACT
Neurones in the optic ganglia of free-moving and fixed Calliphora were electrically stimulated with extracellular microelectrodes. The maximum diameter of the stimulating focus was 20 μ m in the antero-posterior axis. Course control and landing manoeuvres could be elicited by stimulating in the lobula plate but not in the lobula or medulla. With stimulation in the vicinity of the H neurones, yaw responses were evoked. Direction of the response was dependent on stimulation polarity. This supports the hypothesis that H neurones mediate optomotor responses. Stimulation in the posterior part of the lobula plate, close to the V neurones, elicited pitch, lift and thrust responses as well as landing reactions.
INTRODUCTION
The neuronal basis of optomotor responses in insects has been the subject of many investigations (for a partial review see Hausen, 1977). A set of giant neurones in the lobula plate of the fly (Braitenberg, 1972; Pierantoni, 1973; Pierantoni, 1976), have been shown to be sensitive to movement in a large part of the visual field (Bishop & Keehn, 1967; Bishop, Keehn & McCann, 1968; Dvorak, Bishop & Eckert, 1975; Hausen, 1976; Eckert & Bishop, 1978). They make synaptic contact with a small number of large interneurones which descend through the cervical connective to the ventral ganglion. The lobula plate giant neurones thus form part of a link between the chain of visual processing neurones and the motor control centres in the thorax.
The present paper describes the behaviour of free-moving and fixed Calliphora in response to electrical stimulation within the optic ganglia.
MATERIALS AND METHODS
Animals
Experiments were carried out with females of Calliphora erythrocephala, 6-10 days old, from the Institute’s stock. A few control experiments were performed with males within the same range of age and weight. Only flies which showed flight activity in the breeding cages were selected for the experiments. Care was taken to use animals that looked completely healthy, as many flies have damaged wings after being in the breeding cages for a few days.
Room conditions
All experiments were carried out at room temperatures of 19-21 °C and 50-70% relative humidity. Ambient illumination in the room was normally 60-150 lx, and occasionally 5 lx.
Preparation of animals
Flies were briefly anaesthetized with carbon dioxide (for about 5 s) and glued to a cardboard triangle by means of a molten mixture of beeswax and resin or, as a control for the effect of heating with wax, with dental epimine plastic (Scutan). The cemented point was located frontally on the thorax so as to leave the flight movements undisturbed. The fly was fixed to a rack which allowed the head to be pushed down under microscopic control, without damaging the cervical connective. A paper barrier was put around the neck of the fly to prevent the animal from putting its feet in the prepared part of its head. The head was pushed down to an angle of about 70° from the normal position. A small hole (0·5 x 0·5 mm) was made in the back wall of the head just behind the lobula plate, using a broken razor blade. The air sac, which then became visible, was gently pushed to the side, but not removed. The proboscis muscles had to be sectioned, as their contraction induce large movements of the brain. Fat tissue, which sometimes covered the ganglion, was sucked out with a fine tube connected to a vacuum pump. The lobula plate was then easily accessible. A tungsten reference electrode was glued with the wax mixture to a heating wire driven by a micromanipulator and was gently pushed against the lobula plate surface, as close as possible to the position projected for the stimulating electrode, but without penetrating the ganglion. This pressure produced very good mechanical stabilization of the tissue in the vicinity of the stimulated point. Subsequently, the electrode’s shaft was cemented to the head’s wall by a minute drop of Scutan cement and freed from the manipulator after the wax joint has been melted by the heating filament. The tungsten stimulating electrode was pushed just far enough into the tissue to bring the shoulder at the tip (see Fig. 1) into contact with the surface of the ganglion, and was then cemented to the wall of the head. The remaining hole in the head was covered with dental cement. The head was brought back into its normal position and fixed to the thorax with the wax-resin mixture. The wires leading to the electrodes were cemented to the cardboard support. The fly was then fed. After it had rested for one hour, its behaviour was tested. An animal was considered normal if it tried to escape and fly away when attempts were made to catch it from any direction and if no obvious signs of pathological behaviour were detected. The fly was then left undisturbed for a further two hours before an experiment.
Electrodes
Insulated tungsten electrodes (Fig. 1) were used for most experiments. Just before use, the insulation was removed from the electrode tip by sending a short current pulse (1 μA for 10 to 20 ms) through the electrode as its tip was brought into contact with the surface of a bath of either etching fluid or mercury. Mercury was used where the exposed area was to be very small (1 μm or less).
Glass microelectrodes were used for some experiments with fixed flies. The electrodes were filled with physiological saline (modified from Case, 1957) or with 100 mm cobalt chloride solution in water ( + 5% gelatine to avoid leakage if the electrode tip diameter exceeded 2 μm). The electrode tip was finally ground to the required diameter (1 to 10 µm) and bevelling angle (about 45°) on a grinding wheel.
Stimulation
Current stimulation was applied from a constant current source (made in the Institute’s workshop) controlled by a pulse generator (T. Sachs). The intensity of one stimulating current was chosen to be a function of the surface area of the exposed tip. A value of 10 /(μm)2 was never exceeded. Generally, stimulation was made with negative impulses of 0·5 to o·6 ms duration that were spaced by 5-200 ms Stimulation was also attempted with positive impulses and DC current injection of either polarity. To ascertain that the stimulus location coincided with the tip of the electrode, the reference electrode was placed at different positions in the body of the fly. This was repeated for different positions of the stimulating electrode. The reaction of the fly always depended on the postion of the stimulating electrode and not on the position of the reference electrode as long as the total current and current density was kept low (current below 1 μA; current density below 10 μ /(μm)2). Total currents of more than 5 μA elicited traumatic reactions such as convulsions and coma; current density above ionA/(μm)2 produced tissue coagulation in the vicinity of the electrode tip.
Stimulus location
The stimulating electrode was stereo-tactically brought into place. To know the position of the stimulated zone with respect to anatomical landmarks, it was necessary to generate a histologically visible trace of the tip of the electrode. When tungsten needles were used, marking was achieved at the end of the experiment by means of a high frequency current (1 MHz) from a signal generator, which coagulated the tissue in the area of the electrode tip and made the site of stimulation visible in histological sections (Fig. 2 a). When glass microcapillaries were used, the stimulating electrode was filled with cobalt chloride solution. After the stimulation, positive square impulses of 50 ms duration were sent 10 times a second through the electrode for 10 to 20 seconds with an intensity of about 10 nA/(μm)2 of electrode tip area. The brain was rinsed for 10 min with a 1 % ammonium sulphide solution in saline prior to fixation. Fig. 2b shows an example of such a marking experiment.
Histology
The brain was fixed overnight with a 2% glutaraldehyde solution in saline (at pH 7·2), dehydrated in alcohol, embedded in paraffin and cut at 15 μm thickness.
Recording of motor responses
Most of the reported results were obtained from freely moving flies. In this situation, the animal has solely to carry the electrodes and part of the weight of the connecting wires, which altogether and in the worst case amounted to a maximum of 10 mg. The two wires used to connect the stimulating apparatus to the fly were 10 μm in diameter and about 1 m long. The effect of their stiffness on the fly’s behaviour in most cases could be neglected but the fly’s movements were limited by the length of the wires. Responses were recorded with movie cameras (at 18 pictures/s). Event markers in the field of the camera indicated stimulation periods. The position and orientation of the fly was measured frame by frame from the films.
Similar observations were obtained from the stimulation of fixed walking flies in a device kindly provided by E. Buchner (see Buchner, 1976). The fly was fixed in a clamp by the cardboard piece glued to its thorax and its head was tilted down as described earlier. The head was kept in this position for the experiment, and the hole in it was left open to allow electrodes to be inserted repeateadly. The fly walked a small polystyrene ball floating on an air cushion. In this way, electrically stimulated behavioural sequences were largely free of visual feedback. In these experiments no motion pictures were taken and rotations of the ball were not recorded automatically.
RESULTS
Electrical stimulation of the lobula plate was carried out successfully on a total of 48 flies. Twenty-two of them were stimulated with chronic implanted tungsten electrodes; the remaining 26 were stimulated with glass capillaries. A further 8 trials ended unsuccessfully (i.e. no reaction could be elicited at the indicated current levels). No reaction could be evoked that would not fit with one of the types described below-
The effects of electrical stimulation of the lobula plate depended upon position of the stimulation. Three functional layers were found, roughly parallel to the front of the lobula plate.
Stimulation of the frontal third of the ganglion produced very pure yaw manoeuvres, in both flying and walking flies. Direction of the manoeuvres depended on the polarity of the stimulus. Negative current pulses (i.e. the stimulating electrode was the cathode) caused the fly to turn in the direction opposite to the stimulated side. Reversing the stimulus polarity also reversed the direction of the evoked yaw (Figs. 3 and 4). This experiment was extremely repeatable. Similar results were obtained over a wide range of stimulation frequencies (10 to 250 pulses/s with a pulse duration of 0·5 ms); even DC currents evoked yaw. With a fixed fly, the frontal layer of the lobula plate was probed dorso-ventrally and medio-laterally to find the region of highest sensitivity to electrical stimulation. Keeping the stimulation current constant, the intensity of the response appeared to be strongest when the lateral equatorial zone was stimulated, getting weaker the more the stimulus was moved mediad, and no responses or very weak ones could be obtained from the polar regions of the ganglion (Neural projection of zenith and nadir in the visual field). Transient responses of opposite polarity to the reactions described happened in about 10% of stimulations in the frontal layer of the lobula plate. These transients appeared at the beginning and/or at the end of the stimulation and were of short duration (Fig. 5).
Stimulation of the middle third of the lobula plate with negative impulses caused fixed and freely walking flies to walk sideways away from the stimulated side. No similar effect was observed during flight. The walking reaction could be evoked over a wide field in the frontal plane (i.e. medio-laterally and dorso-ventrally). The most effective region for stimulation was found to be in the middle of the lobula plate. The response could only be obtained with trains of negative impulses (0·5 ms duration and at rates of 50 to 200 impulses/s). D.C. currents and positive impulses failed to elicit the reaction.
Stimulation of the caudal third of the lobula plate evoked several reactions which all involved manoeuvres in the sagittal plane (stimuli were negative impulses with a rate of 50 to 200 impulses/s).
Stimulation of the ventral middle third of this layer induced a landing reaction in flight (defined as a spreading of all legs without interruption of the wing beat).,This landing reaction was paired with the return of the antennae to a more or less vertical position (in flight, the antennae are thrust forward). A somewhat similar reaction was elicited when the electrode was inserted into the lateral border of the ganglion, but in this situation only the front legs were spread forward (see Eckert & Bishop, 1978). This last reaction could, surprisingly, also be evoked in the walking fly, which seemed to try to catch something in front of itself and above its head.
Stimulation of the lateral half of the caudal layer evoked upward pitch in flight.
Stimulation of the medial half of the layer caused the walking fly to lower its body onto the ground. In free-flying individuals, stimulation in this area caused an increase in lift and thrust together with upward pitch. This reaction started relatively late after the beginning of the stimulation. Fig. 6 is a rough summary of all the results.
DISCUSSION
The elicited behavioural responses had a co-ordinated appearance. It should be emphasized that at low current densities confused motor activity like tumbling, convulsion, leg shaking or rivalry between behavioural activities was never observed. When the stimuli were applied during flight, the forces exerted by the animal did not decrease. When walking animals were stimulated, the evoked movements were fluent sequences, even when the behaviour evoked appeared but rarely in the unstimulated fly. All evoked behaviour was, however, observed in free-moving flies of the absence of any artificial stimulation. It thus seems that the electrically evoked behaviour patterns closely mimicked those existing naturally.
In the present study, care was taken to keep the focus of stimulation as small as possible. Electrode tips were made small to concentrate the stimulating current, and currents were made small to reduce the probability that low resistive channels somewhere in the neuropile could concentrate the current lines sufficiently to generate a ‘ghost’ stimulation point (Huber, 1960). No precise measurement of the range of the stimulation could be made, but after a response was elicited, a rough indication was provided by the distance the electrode had to be pushed forwards before the response disappeared. This distance was 5 to 7 μm with electrodes of less than 2 μm of exposed tip. A smaller value would be unlikely to be obtained since the range of mechanical instability of the preparation was 5 μm.
In comparing responses observed in this study with the available data for the lobula plate neurones, due consideration should be made of the experiments situation. In the present experiments, the artificial commands were superimposed on the flies endogenous movements. Furthermore, the evoked manoeuvres caused visual feedback which was probably not suppressed by efference copy. The latter problem could be eliminated using the apparatus described in the next paper (Blondeau, 1981). The problem was partly overcome in this study with fixed flies walking on a styrofoam ball. In future experiments, white-eyed flies with ‘neutralized’ corneas (Francescini & Kirschfeld, 1971) might be used to achieve satisfactory open loop conditions, while avoiding complete darkness.
Production of yaw by stimulation of the anterior region of the lobula plate (see Fig. 7 for a general sketch of the lobula plate’s architecture) is in accord with the presence of H neurones in this region (Pierantoni, 1973). However, the direction of the yaw response indicates a more complicated effect than just stimulation of the H neurones. Turns away from the stimulated side were made when the extracellular stimulation was made with negative pulses, which should depolarize nearby cells, Depolarization of H neurones should produce turns towards the stimulated side (Hausen, 1976). It should be noted, however, that transient responses of opposite Polarity were occasionally observed at the beginning and end of a stimulation period.
Current applied to the middle layer of the lobula plate can be expected stimulate the V(2) neurones (Hausen, 1976). These neurones are depolarized by upwards as well as back to front movements in the ipsilateral visual field (Hausen, 1976). The exact connection of V(2) cells are still unknown, but it seems probable that they, too, make direct contact to descending neurones (Hausen, 1976). The nature of the sideways walk that could be elicited in this medial layer is too little understood to make a comparison with the properties of the V(2) cells meaningful (see also Eckert & Bishop, 1978). If this walking was caused by a transient roll reaction (counteracted by the autonomous balance mechanism of the thorax) its polarity would be compatible with a direct effect upon the V(2) neurone (Hausen, 1976).
Pitch, lift and thrust were elicited by stimulation of the posterior layer of the lobula plate. V and V(1) neurones which cover the posterior surface of the lobula plate are depolarized by ipsilateral downward movement (Hausen, 1976; Eckert & Bishop, 1978). As in the case of the electrically stimulated yaw response, the polarity of the pitch response is the opposite of what would be expected if the negative stimulus were to depolarize V neurones. Recent behavioural studies on a mutant of Drosophila (J. Blondeau, in preparation) indicate that V cells are involved in visua control of pitch and roll. Optomotor lift and thrust responses seem to be independent of V neurones in Drosophila (Heisenberg, Wonneberger & Wolf, 1978). The excitability of these responses in the posterior layer of the lobula plate in Calliphora might suggest the existence of other neurones in this layer which are involved in altitude control, but the following explanation is at least as plausible. If one keeps in mind that in walking flies forward movement could never be elicited by electrical stimulation, the late appearance of the increase in lift and thrust after the onset of the stimulus in flight may indicate that these forces are a secondary effect of the pitch response in order to avoid stall and loss of altitude.
ACKNOWLEDGEMENTS
In summary, the agreement between the properties of neurones in the lobula plate and the behavioural responses elicited by weak electrical stimulation in their vicinity indicates the key role that individual neurones seem to play in invertebrate behaviour.